- Email: [email protected]
Monkey insular cortex neurons respond to baroreceptive and somatosensory convergent inputs
Convergent baroreceptor and somatic neurons in monkey insular cortexNeuroscience Vol. 94, No. 2, pp. 351–360, 1999 351
Pergamon PII: S0306-4522(99)00339-5
Copyright q 1999 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306-4522/99 $20.00+0.00
MONKEY INSULAR CORTEX NEURONS RESPOND TO BARORECEPTIVE AND SOMATOSENSORY CONVERGENT INPUTS Z.-H. ZHANG,* P. M. DOUGHERTY† and S. M. OPPENHEIMER*‡ *Laboratory of Neurocardiology, Cerebrovascular Division, Department of Neurology and †Departments of Neurosurgery and Neuroscience, The Johns Hopkins University School of Medicine, Baltimore, MD 21287-7585, U.S.A.
Abstract—To investigate possible convergence of autonomic and somatosensory input in the insula of the non-human primate, extracellular single-unit recordings were obtained from 81 neurons (43 insular and 38 in surrounding cortex) during application of cutaneous nociceptive stimuli (pinch) and baroreceptor challenge in six anesthetized monkeys (Macaca fascicularis). All cells were also tested with light touch (brush) stimulation. Twenty-six units responded to blood pressure changes; 20 (80%) were identified within the insula (P , 0.001). The majority of these insular units (16/20) also responded to nociceptive pinch (convergent units). More units responsive to changes in blood pressure (unimodal and convergent) were found in the right (18/29, 62%) than in the left insular cortex (2/14, 14%)(P 0.004). Twenty-nine insular neurons responded to nociceptive stimuli; 16 of these were convergent units and 13 showed unimodal responses to somatosensory stimuli alone. These cells had wide bilateral receptive fields including face, hand, foot and tail. Ten insular neurons were unresponsive to both sets of stimuli (non-responsive cells); significantly more of these cells (28/38) were identified in extrainsular locations (P , 0.01). We suggest that the primate insular cortex may be involved in the integration of cardiovascular function with somatosensory (principally nociceptive) input. This view supports the emerging role of the insular cortex as an important forebrain site of viscerosomatosensory regulation with clinical implications for cardiovascular regulation under conditions of stress and arousal. q 1999 IBRO. Published by Elsevier Science Ltd. Key words: extracellular recordings, insular cortex, pain, cardiovascular regulation.
the Study of Pain and the NIH guide for the care and use of laboratory animals. Six adult male monkeys (Macaca fascicularis) weighing 4.5–7.5 kg were tranquilized with ketamine (10 mg/kg, i.m.) during transportation to the laboratory and catheterization. The right radial artery was cannulated with PE-90 tubing filled with heparinized saline (10 units/ml) for blood pressure recording. The right cephalic vein was cannulated with an i.v. catheter for infusion of anesthetics and drugs. Surgical anesthesia was induced by an intravenous dose of pentobarbital (25 mg/kg) and later maintained by intravenous infusion of pentobarbital (5 mg/kg/h). Once anesthesia was deep, the monkeys were paralysed with pancuronium bromide (0.1 mg/kg) and artificially ventilated. Throughout the surgical procedures the level of anesthesia was periodically tested and adjusted as necessary by examining pupillary size, and continuously monitored by recording expired CO2, blood pressure and heart rate. End-tidal CO2 was kept between 3.5% and 4.5% by adjustment of respiration rate and depth, and core temperature was maintained at 37–388C by a servo-controlled heating blanket. The animals were placed in a stereotaxic frame and a partial craniectomy was made on both sides for exposure of the insular cortex.
In many species, the insular cortex is an important cortical site of cardiovascular representation. 2,10,25,29,34,41 In the rat, both electrical and chemical stimulation of the insula caudal to the crossing of the anterior commissure produces cardiac chronotropic responses and blood pressure changes. 25,37 Prolonged insular stimulation in rats generates lethal cardiac arrhythmias and structural changes in cardiac muscle. 26 Previous electrical stimulation studies in the monkey have suggested similar insular involvement in cardiovascular control but lacked the precision of contemporary methods. 20,35 Anatomical evidence in primates has implicated the insula in the processing of somatosensory innocuous and nociceptive information. 17 In addition, functional imaging studies in humans using positron emission tomography have demonstrated insular activation with somatosensory (especially thermal and vibrotactile) stimuli. 8,13,15 The goal of this study was to investigate whether primate baroreceptor responsive insular neurons also respond to somatic stimuli, as has recently been shown in the rat insular cortex 19 and thalamus. 40 Such convergence at cortical levels could further support the emerging role of the insula as a site of integration of pain and autonomic function.
Recording procedures Extracellular recordings were made from spontaneously firing neurons in the insular cortex and surrounding regions with carbon filament-glass microelectrodes (impedance 1–2 MV at 1 kHz). Coordinates for the regions of study (AP 11–20 mm; ML 15–18 mm; DEPTH 10–20 mm from cortical surface) were based on the stereotaxic atlas of Szabo and Cowan 31 adjusted for body weight. Singleunit activity, electrocardiogram and arterial pressure were amplified and displayed on an oscilloscope (Hitachi VC 6155, Japan) and fed to a computer running a neurophysiological data acquisition and analysis package (Datawave Systems, Discovery package, Version 4.0, Thornton, CO, U.S.A.). Spike size and configuration were continuously monitored to confirm that activity of the same cell was recorded throughout the experiment. Single-unit activity was discriminated from background and stored on computer for later off-line analysis.
EXPERIMENTAL PROCEDURES
Surgical procedures All procedures were approved by the Institutional Animal Care and Use Committee of the Johns Hopkins University and are consistent with the guidelines of the International Association for ‡To whom correspondence should be addressed. Abbreviations: PB, pinch and brush; PE, phenylephrine hydrochloride; PET, positron emission tomography; PR, pinch responsive; SE, sympathoexcitatory; SI, sympathoinhibitory; SNP, sodium nitroprusside; VPL, ventroposterior lateral nucleus.
Baroreceptor challenge and cutaneous stimuli Following unit isolation and stabilization, 3 min of spontaneous 351
352
Z.-H. Zhang et al.
the recording sites were identified with reference to a measured scale taking into account the magnification and the distance between the recording site and the bottom of the microelectrode track. Correction for brain shrinkage was made to ensure precise histology by measuring the depth of the track in the pre- and postfixed states. Tissue distortion interfering with track reconstruction was rarely encountered. Significant brain edema was not generally seen. In all cases (except one) the entire track could be seen in at least one section, making reconstruction feasible (Fig. 1). However, in one monkey, tissue was distorted owing to brain edema. Identification of the bottom of the track was not impaired and correction for the distortion was possible by careful reconstruction of the track in sequential sections. Data analysis The stored digital records of unit activity before and after baroreceptor challenge and somatosensory stimulation were retrieved and analysed off-line. Baroreceptive responses were analysed by comparison of discharge rates for 90–120 s after drug injection compared with a similar period before injection. Mechanoreceptor responses were analysed by comparison of discharge rates in each separate 10 s segment of stimulus application to a 10 s segment of pre-stimulus activity. All values were expressed as the mean ^ S.E.M. Changes in firing rates for individual cells . ^ 20% were considered significant. 1,38,39,41 Cells with such changes were then categorized into groups according to their responses to PE, SNP, pinch and brush. Changes in their mean firing rate for the periods before and during challenge with baroreceptor and somatosensory stimuli were analysed statistically using Student’s t-test. The distributions of responsive neurons were assessed for intrainsular or extrainsular localization and lateralization with Fisher’s exact test. P , 0.05 indicated statistical significance. Fig. 1. Photomicrograph of a Thionine-stained section showing the track (indicated by the arrow) through the dysgranular responsive laminae of the insular cortex. Scale bar 5 mm.
baseline activity was recorded. Arterial baroreceptors were then challenged by bolus intravenous administration of the pressor agent phenylephrine hydrochloride (PE, 10 mg/kg, Sigma Chemical Co., St Louis, U.S.A.) and the depressor drug sodium nitroprusside (SNP, 5 mg/kg, Sigma Chemical Co., St Louis, U.S.A.) in a volume of 0.2 ml 0.9% saline, followed by 0.2 ml 0.9% saline to flush the dead space of the tubing (dead space volume 0.1 ml). These doses induced ^20–40 mmHg change in arterial pressure. Injection of 0.2 ml 0.9% saline was used as a control procedure. These injections were given in random sequence. After characterization of unit responses to baroreceptor activation and re-establishment of baseline firing conditions, a further 2 min of spontaneous baseline firing activity was collected. The responses of cells to a series of mechanical stimuli applied to the selected skin sites were then determined. Cutaneous mechanoreceptor stimuli included application of a camel hair brush and arterial clip to seven areas of the body (cheeks, hands, feet and tail). Clip pressure was standardized using a spring mechanism. When applied to the human skin, the clip elicited pain in a reproducible fashion. Each stimulus set began with recording of background activity followed by the brush (back and forth brushing) applied to site 1 (randomized) for 10 s. Following a 10–20 s interval, or until baseline firing rate was re-established following removal of the stimulus, the brush was then applied to site 2. This continued until the brush had been applied to all sites. The same sequence was then repeated for the pinch stimulus. Care was taken to ensure that the brush and pinch stimuli were delivered in a stereotyped manner on each occasion. Histology At the conclusion of each experiment, the last recording site was marked by the insertion of a steel electrode for placement of an electrolytic lesion (100 mA d.c. anodal current, 10 s duration). Other cell sites were identified with respect to this reference point. The animals were killed by an overdose of pentobarbital, and the brain was removed and postfixed in 10% formalin for at least 10 days. Frontal sections of 50 mm thickness were cut with a cryostat microtome. Sections were mounted on gelatin-coated glass slides, dried overnight, dehydrated in a series of graded alcohols and then stained with 0.125% Thionin solution. The slide was projected using a Zeiss camera lucida, and
RESULTS
Neuronal responses to baroreceptor and mechanoreceptor stimuli were investigated in 81 units, 43 within the insula and 38 in adjacent cortex. As the investigation was primarily designed to investigate the presence of convergent cardiovascular-nociceptive cells within the insula, no attempt was made to investigate subcortical areas including the claustrum and basal ganglia. The insular cortex was defined as that part of the cerebrum that overlies the claustrum. In primates this region lies buried beneath the frontoparietal and temporal opercula. Three cytoarchitectural divisions or fields of the insula occur in the monkey. 2,21 These include a rostroventral agranular field, a transitional dysgranular field and, posteriorly, a granular field. Cells were categorized in accordance with their firing responses to baroreceptor activation by PE and SNP-induced blood pressure changes 3,39,41 and by responses to somatic stimuli. 1,18,36 Convergent units responded to both blood pressure changes and somatic stimulation. Neuronal response patterns to baroreceptor activation A total of 26 units (20 in the insula and six in the adjacent cortex) responded to baroreceptor challenge with PE and SNP. According to their response to PE and SNP, two types of unit were categorized as reported previously in the rat 3,9,39 and monkey. 41 Sympathoexcitatory (SE) units were identified by a significantly decreased firing rate following PE-induced increase in blood pressure. The firing rate of these units increased with SNP-related reduction in blood pressure (Fig. 2A, B). Sympathoinhibitory (SI) units showed an increased firing rate with PE injection and decreased firing following SNP administration (Fig. 3A, B). The classification of cells as either SE or SI units is a convention 3,39,41 and does not necessarily mean that the physiological effect of these cells is either to increase or decrease cardiovascular
Fig. 2. Firing patterns of a representative SE nociceptive convergent unit in the monkey insula. Firing rate decreases with PE (10 mg/kg) induced blood pressure rises (A) and increases with SNP (5 mg/kg) induced falls in blood pressure (B). Panel (C) shows absence of response to the application of brush stimulation at seven sites. Application of pinch to these seven sites increases discharge rate (D). In each panel: Top trace shows blood pressure (BP); middle trace shows the footpad marker of seven brush (C) and pinch (D) sites; bottom trace shows the frequency histogram of neural firing (binwidth 1 s). The right column shows the superimposed traces of neural spikes showing consistency of amplitude and shape throughout the recording. Asterisks show significant increases in firing rate during pinch stimuli; the corresponding responsive sites are marked by black dots on the monkey diagram (D). The arrows indicate drug administration (A and B).
354
Z.-H. Zhang et al.
Fig. 3. A representative response of a SI nociceptive convergent unit in monkey insular cortex showing firing rate increases after PE (10 mg/kg) injection (A) and decreases following SNP (5 mg/kg) administration (B). No significant response to brush correlating with stimulus application (C). Significant increases in firing rate to pinch in all seven sites (D). Display and abbreviations are the same as for Fig. 1.
Convergent baroreceptor and somatic neurons in monkey insular cortex
sympathetic tone. Non-responsive units showed no response to either baroreceptor or somatic activation (Fig. 4). Equal volume injections of saline had no effect on discharge rates in these same cells. The majority of insular cortex responsive cells showed SE responses to blood pressure changes (14/20, 70%) (Table 1). SI responses (6/20, 30%) were relatively infrequently encountered in the insula (Table 1). The mean firing rates of SE and SI units are shown in Fig. 5. For SE units the mean onset latency was 7.9 ^ 2.6 s (range 0–20 s) for PE injections and 10.4 ^ 3.2 s (range 0–22 s) for SNP injections. The mean duration of the SE response to PE injections was 102 ^ 14 s (range 20–180 s), and for SNP the mean duration of the response was 94 ^ 13 s (range 20– 160 s). For SI units, the mean onset latency for PE injections was 8.6 ^ 2.4 s (range 0–15 s) and 11.2 ^ 2.7 s (range 0–18s) for SNP injections. The response duration to PE was 94 ^ 16 s (range 20–150 s) and 97 ^ 16 s (range 30–170 s) for SNP. Neuronal response patterns to somatic stimuli Thirty-seven units (29 in the insula and eight outside the insula) showed a significant change in firing rate with mechanical stimulation of the skin. Responses showed clear onset and offset following the 10 s application of somatic stimuli. The mean firing frequency before and during somatic stimulation for each group is shown in Fig. 5. Most cells (n 31) showed increased discharges to nociceptive stimulation (on-cell) while the remainder (n 6) showed decreased firing (off-cell). Thirty-one showed responses to pinch only [26 on-cells, five off-cells; pinchresponsive (PR) units]; six units responded to both pinch and brush (PB units); no units were observed to respond to brush alone (Fig. 5, Table 2). Nociceptive units (PR and PB cells) had relatively wide receptive fields: 19/37 (51%) responded to stimulation applied to five or more sites throughout the upper and lower body; 6/37 (16%) cells responded to contralateral somatic sites. Five cells responded to bilateral “upper” body stimuli and three cells to bilateral “lower” body stimuli. Four neurons responded only to ipsilateral stimuli applied to more than one site. Pinch induced a very slight decrease in blood pressure (,10 mmHg) but a marked increase in firing rate (Figs 2, 3). These blood pressure responses were highly variable in their occurrence (Figs 2, 3). The effect of SNP injection was to induce a much greater fall in blood pressure in SE cells than that which occurred with pinch, but this was accompanied by a much smaller increase in firing rate. For SI cells, where pinch induced a small change in blood pressure, the effect on neural firing was opposite to that induced by SNP (Fig. 3). Categorization of convergent neuronal responses Twenty units responded to both baroreceptor challenge and somatic mechanical stimuli (convergent units). Sixteen of 29 insular cells were convergent units; four extrainsular cells showed such convergent responses. Most of the insular convergent units (11/16, 69%) showed sympathoexcitatory responses to baroreceptor activation (Table 2, Fig. 6). Five SI convergent units were found within the insular cortex (Table 2, Fig. 6). Fifteen of 20 convergent cells responded
355
only to pinch stimuli (PR cells); 5/20 convergent units responded to both pinch and brush stimulation (PB cells); no convergent units showed responses to brush alone (Table 2, Fig. 5). Thirty-eight non-responsive neurons responded to neither baroreceptor challenge nor somatic stimuli (Table 2, Figs 5, 6). Localization of responsive and non-responsive units The majority of insular neurons (33/43, 77%) responded to baroreceptor and/or somatic activation. Fewer such responses were seen in the adjacent cortex (10/38, 26%) (Fig. 6, Table 1). More units responsive to changes in blood pressure (unimodal and convergent) were found in the right (18/29, 62%) than in the left insular cortex (2/14, 14%)(P 0.004, Fig. 6, Table 1). There was no clear localization of responsive cell types in the rostrocaudal insular axis. A significantly greater number of cells unresponsive to either blood pressure changes or cutaneous mechanoreceptor stimuli was found in extrainsular (28/38, 74%) compared with intrainsular locations (10/43, 23%) (P , 0.01, Table 1, Fig. 6). Responsive insular neurons were chiefly identified within the midinsular dysgranular region in layers II, III and V/VI (Figs 1, 6). Few responsive units were identified in the agranular insular cortex. This may reflect the limited sampling of the agranular region. More anterior and posterior insular regions were not systematically sampled. Responsive neurons outside the insula were encountered in the left mid-frontoparietal operculum adjacent to the insula (two convergent, two SI and one PR unit) and left temporal operculum (one PB unit), and right superior/mid-temporal cortex (two convergent and two PR units) (Fig. 6, Table 1). DISCUSSION
Response patterns of insular convergent neurons This study shows that cells convergent for baroreceptive and somatosensory input can be identified within the insular cortex of the anesthetized monkey. As previously reported in the rat and monkey, the majority of insular neurons responsive to blood pressure changes were SE units. 39,41 Convergent units generally responded to nociceptive stimuli. The data showed the possibility of lateralization of units responsive to blood pressure changes within the primate insula. However, this may reflect sampling bias as fewer left anterior insular regions were sampled in this study. The level at which convergence of baroreceptive and nociceptive input occurs initially within the CNS cannot be determined by this study. However, we have recently demonstrated such convergence within the posterior thalamus of the rat in units which project monosynaptically to the insula, 40 suggesting that convergence is likely to occur at subcortical sites in the monkey. Visceral (non-cardiovascular) and somatic convergent neurons have been demonstrated within the ventroposterior lateral nucleus (VPL) of the monkey 4 and cat thalamus. 32 A recent study suggested that an inhibitory interaction occurs between visceral and somatic inputs in the thalamic VPL of the squirrel monkey. 5 This implies that either somatic conditioning stimuli inhibit visceral responses, or vice versa. Similar interactions have been identified at different levels of the CNS in other species; 11,12,23,24 however, this has not yet been demonstrated in the primate insular cortex. We believe for several reasons that a distinction can be
356
Z.-H. Zhang et al.
Fig. 4. Representation of a non-responsive unit in the extrainsular frontoparietal operculum showing no significant change in firing rate following blood pressure alteration by PE and SNP (A and B). No response to brush and pinch stimuli (C, D). Display and abbreviations are the same as for Fig. 1.
Convergent baroreceptor and somatic neurons in monkey insular cortex Table 1. Distribution of baroreceptive, somatosensory, convergent and nonresponsive cells within and outside the insula
SE cell SI cell PR and PB cell Convergent cell Non-responsive cell Total
Right in IC
Left in IC
Right out IC
Left out IC
Total
3 1 7 14 4 29
0 0 6 2 6 14
0 0 2 2 13 17
0 2 2 2 15 21
3 3 17 20 38 81
SE and SI cells respond only to baroreceptor stimulation with PE and SNP. PR cells respond to pinch only and PB cells respond to pinch and brush. Non-responsive cells respond to neither baroreceptor nor pinch and brush stimulation. IC, insular cortex.
made in the firing of convergent insular neurons to nociceptive stimuli. The responses are not simply due to the slight changes in blood pressure induced by pinch. Firstly, the nociceptive responses have a clear onset and offset with application of pinch stimuli and do not correspond to the somewhat delayed blood pressure responses to pinch (Figs 2, 3). Secondly, the majority of cells showed an increase in firing rate with SNP-induced blood pressure decrease (SE cells). Although pinch also induced a decrease in blood pressure, this was minimal compared with the SNP-induced blood pressure changes, but was associated with a much greater increase in firing than achieved by the larger SNP-induced decrease in blood pressure (Fig. 2). In the case of SI cells, pinch induced a decrease in blood pressure but an increase in firing, opposite to the effects of SNP on the cell (Fig. 3). Finally, blood pressure responses to pinch were inconsistent; responses to pinch were often obtained in the absence of any blood pressure effect (Figs 2, 3). Whether convergent and non-responsive cells in the
357
Table 2. Classification of baroreceptive and somatosensory units Somatosensory responses
PR PB NR Total
Baroreceptive responses SE SI NR 11 3 3 17
4 2 3 9
16 1 38 55
Total 31 6 44 81
NR, non-responsive units.
primate insula respond to other sensory modalities such as taste or gastric mechanical input, as has been shown for rat anterior insular neurons by Cechetto and Saper, 9 was not the objective of this investigation, which focussed on the demonstration of nociceptive–baroreceptive convergence. In theory, specificity of neuronal response to baroreceptor input may be confirmed by baroreceptor denervation. This would counter arguments that responses to blood pressure manipulations were produced by stimulation of mechanoreceptors in the arterial wall or adjacent structures rather than by stimulation of baroreceptors in the aortic arch, heart and carotid sinus. In practice, however, baroreceptor denervation in the monkey also produces an unstable preparation. The cardiovascular status of the animals in these two widely differing situations (pre- and post-denervation) would no longer be comparable, confounding comparison of cellular responses between these states. We argue that there is specificity of baroreceptor response on the basis of the reproducible manner in which these cells responded to standard techniques of baroreceptor activation with PE and SNP, as we have already shown for insular baroreceptive neurons in the rat 39 and monkey. 41 Other studies have indicated that similar stimuli induce their effect on units in the brain by activation of peripheral visceral receptors. 9
Fig. 5. Statistical analysis of firing rates of cells divided into groups according to their responses to baroreceptor and mechanical stimulation. Data are shown for all 81 cells in the study. SE convergent cells are characterized as in Fig. 1; SI convergent cells are defined as in Fig. 2. PR and PB cells respond to pinch (PR cells) or pinch and brush stimuli (PB cells) but not to baroreceptor challenge with PE and SNP; SE and SI cells respond only to baroreceptor challenge with PE and SNP (as defined in Figs 1 and 2) but not to somatosensory stimuli. Statistical analysis was conducted for changes in unit firing rates to somatosensory stimuli application (brush and pinch) at all responsive sites. On or off cell responses to pinch within each group are shown separately (Pinch-on; pinch-off). No off responses to brush stimulation were encountered. No statistical testing was conducted on unimodal SE (n 3) and SI cells (n 3) in response to SNP and PE challenge, because of the small number of cells with these characteristics. In each group, the hatched bars represent baseline firing rates before drug injection or mechanoreceptor stimuli (control); the solid bars reflect the changes in firing rate after drug infusion or during mechanical stimulation. *P , 0.05; **P , 0.01; ***P , 0.001 (Student’s t-test for paired data).
358
Z.-H. Zhang et al.
Functional organization of the monkey insular cortex Cardiovascular responses have previously been shown in earlier stimulation studies of the monkey insula. 20,35 However,
care must be taken in the interpretation of these data because responses were elicited with large constant voltage stimuli of long duration (2–10 ms) using large surface stimulating electrodes. Consequently, the specificity of the responses and their localization to the stimulation site may be questionable. Previous studies indicated that approximately 70% of neurons in the granular posterior insular cortex respond to non-noxious somatic inputs in anesthetized cynomolgus monkeys. 27,28 Only 24 neurons were investigated in this region, of which 25% also responded to nociceptive stimuli. We did not identify any cells that responded solely to nonnociceptive stimuli. However, we explored the dysgranular mid-insula and did not sample the very posterior regions of the insular cortex studied by Robinson and Burton. 27,28 This methodological difference may explain this discrepancy. Sixty per cent of insular neurons in the Robinson and Burton study responded to bilateral somatic stimulation, consistent with our findings for PR and PB cells. Similar results have been found in unanesthetized macaques. 30 The monkey insula receives projections from the posterior part of the thalamic ventral medial nucleus, 14 which in turn receives input from spinal dorsal horn lamina I neurons, a site of termination of nociceptive primary afferents. Burkey et al. 6 recently found that microinjection of morphine into an anatomically defined opioid-responsive site in the rostral agranular insular cortex of the rat significantly reduced the behavioral response to noxious stimuli. Their more recent data 7 indicated that a descending tonic inhibitory dopaminergic pathway from the insula can modulate nociception. Anatomical studies suggest from a comparative consideration of connectivity that the rostral agranular insular cortex in the rat is homologous to more posterior areas in the primate insular cortex which have been implicated in the development of post-stroke pain syndromes. 33 It is worthy of note that we found that some cells in the area adjoining the insular cortex also receive baroreceptor and/or nociceptive inputs. No overall conclusion can be reached, owing to limited sampling of these areas. Recent functional imaging studies show involvement of the frontoparietal operculum in nociceptive processing in the rat. 22 Our studies in the rat and monkey also showed that a small percentage of neurons adjacent to the insular cortex (most in the parietal operculum) responds to blood pressure change. 39,41 However, further studies are needed to clarify this issue. Nociceptive input into the insular cortex has also been shown using positron emission tomography (PET). Increased insular cerebral blood flow occurred in patients complaining of atypical facial pain when painful but not when non-painful thermal stimulation was applied to the back of the contralateral hand. 16 Interestingly, normal controls did not show this increased insular activation, suggesting the involvement of the insula in the affective component of pain. Likewise, functional imaging studies in humans show strong activation of the mid/anterior insula by noxious and innocuous thermal Fig. 6. Distribution of convergent (B), SE (X), SI (K), nociceptive (V), and non-responsive (W) units within the monkey insula and surrounding regions. Numbers indicate the distance in mm anterior to the interaural zero point. Amyg, amygdala; CA, anterior commissure; Cd, caudate nucleus; Cho, optic chiasm; Cl, claustrum; Gp, globus pallidus; Hyp, hypothalamus; Put, putamen; TO, optic tract; VA, ventral anterior nucleus of the thalamus; VLm, ventral lateral nucleus of the thalamus, medial part; VLo, ventral lateral nucleus of the thalamus, oral part. Figure templates are modified from the atlas of Szabo and Cowan. 31
Convergent baroreceptor and somatic neurons in monkey insular cortex
stimuli. 8,15 It has been hypothesized by these investigators that the insula is concerned with the conscious perception and discrimination of the affective qualities of both thermal and noxious stimuli and may be involved in the generation of illusory centrally mediated sensations such as phantom limb pain. Coghill and colleagues 13 used PET to localize differential insular activation with vibrotactile stimulation compared with noxious thermal stimuli applied to the left forearm. Nociceptive activation of the insula was noted in the principally contralateral anterior regions, although some ipsilateral responses were apparent. Vibrotactile responses were principally confined to the posterior insula. As in our study, they concluded that there was significant input of nociceptive information into the insula and suggested possible involvement of the dysgranular insula consistent with our findings. However, we did not find cells solely responsive to non-nociceptive inputs. This discrepancy may relate to the use of
359
anesthesia in this study, species differences and the different stimuli used to elicit tactile responses. CONCLUSIONS
Our data indicate significant convergence of somatosensory (principally nociceptive) and baroreceptive input on the same neurons in the primate insula. This region has recently been shown to be involved with the affective and perceptive elements of pain processing in human functional imaging studies. There is much recent evidence indicating that the insula also has an important role in the efferent control of cardiovascular function. Taken together, these data indicate that the insular cortex may co-ordinate the autonomic aspects of central pain processing preparing the animal for appropriate enhancement or suppression of cardiovascular function as the environmental circumstances dictate.
REFERENCES
1. Andersen E. and Dafny N. (1983) Dorsal raphe stimulation reduces responses of parafascicular neurons to noxious stimulation. Pain 15, 323–331. 2. Augustine J. R. (1996) Circuitry and functional aspects of the insular lobe in primates including humans. Brain Res. Rev. 22, 229–244. 3. Barman S. M. (1990) Descending projections of hypothalamic neurons with sympathetic nerve-related activity. J. Neurophysiol. 64, 1019–1032. 4. Bruggemann J., Shi T. and Apkarian A. V. (1994) Squirrel monkey lateral thalamus. II. Viscerosomatic convergent representation of urinary bladder, colon, and esophagus. J. Neurosci. 14, 6796–6814. 5. Bruggemann J., Shi T. and Apkarian A. V. (1998) Viscerosomatic interactions in the thalamic ventral posterolateral nucleus (VPL) of the squirrel monkey. Brain Res. 787, 269–276. 6. Burkey A. R., Carstens E., Wenniger J. J., Tang J. W. and Jasmin L. (1996) An opioidergic cortical antinociception triggering site in the agranular insular cortex of the rat that contributes to morphine antinociception. J. Neurosci. 16, 6612–6623. 7. Burkey A. R., Carstens E., Wenniger J. J. and Jasmin L. (1997) Antinociception and descending inhibition from dopamine in the rostral agranular insular cortex of the rat. Soc. Neurosci. Abstr. 23, 160. 8. Casey K. L., Minoshima S., Morrow T. J. and Koeppe R. A. (1996) Comparison of human cerebral activation patterns during cutaneous warmth, heat pain, and deep cold pain. J. Neurophysiol. 76, 571–581. 9. Cechetto D. F. and Saper C. B. (1987) Evidence for a viscerotopic sensory representation in the cortex and thalamus in the rat. J. comp. Neurol. 262, 27–45. 10. Cechetto D. F. (1994) Identification of a cortical site for stress-induced cardiovascular dysfunction. Integr. physiol. behav. Sci. 29, 362–373. 11. Chernigovsky V. N., Musyashchikova S. S. and Sinyaya M. S. (1978) Interaction of visceral and somatic afferent systems in the cerebral cortex. Neurosci. behav. Physiol. 9, 273–282. 12. Coffin B., Azpiroz F. and Malagelada J. R. (1994) Somatic stimulation reduces perception of gut distention in humans. Gastroenterology 107, 1636–1642. 13. Coghill R. C., Talbot J. D., Evans A. C., Meyer E., Gjedde A., Bushell M. C. and Duncan G. H. (1994) Distributed processing of pain and vibration by the human brain. J. Neurosci. 14, 4095–4108. 14. Craig A. D. (1996) Supraspinal projections of lamina I neurons. In Forebrain Processing of Pain (ed. Besson J.-M.), pp. 13–25. John Libbey, Paris. 15. Craig A. D., Reiman E. M., Evans A. and Bushnell M. C. (1996) Functional imaging of an illusion of pain. Nature 384, 258–260. 16. Derbyshire S. W. G., Jones A. K. P., Devani P., Friston K. J., Feinmann C., Harris M., Pearce S., Watson J. D. G. and Frackowiak R. S. J. (1994) Cerebral responses to pain in patients with atypical facial pain measured by positron emission tomography. J. Neurol. Neurosurg. Psychiat. 57, 1166–1172. 17. Dostrovsky J. O. and Craig A. D. (1996) Nociceptive neurons in primate insular cortex. Soc. Neurosci. Abstr. 22, 111. 18. Dougherty P. M., Palecek J., Paleckova V., Sorkin L. S. and Willis W. D. (1992) The role of NMDA and non-NMDA excitatory amino acid receptors in the excitation of primate spinothalamic tract neurons by mechanical, chemical, thermal, and electrical stimuli. J. Neurosci. 12, 3025–3041. 19. Hanamori T., Kunitake T. and Kannan H. (1998) Responses of neurons in the insular cortex to gustatory, visceral, and nociceptive stimuli in rats. J. Neurophysiol. 79, 2535–2545. 20. Hoffman B. L. and Rasmussen T. (1953) Stimulation studies of insular cortex of Macaca mulatta. J. Neurophysiol. 16, 343–351. 21. Mesulam M.-M. and Mufson E. J. (1982) Insula of the old world monkey. I: Architectonics in the insulo-orbito-temporal component of the paralimbic brain. J. comp. Neurol. 212, 1–22. 22. Morrow T. J., Paulson P. E., Danneman P. J. and Casey K. L. (1998) Regional changes in forebrain activation during the early and late phase of formalin nociception: analysis using cerebral blood flow in the rat. Pain 75, 355–365. 23. Ness T. J. and Gebhart G. F. (1991) Interactions between visceral and cutaneous nociception in the rat. I. Noxious cutaneous stimuli inhibit visceral nociceptive neurons and reflexes. J. Neurophysiol. 66, 20–28. 24. Ness T. J. and Gebhart G. F. (1991) Interactions between visceral and cutaneous nociception in the rat. II. Noxious visceral stimuli inhibit cutaneous nociceptive neurons and reflexes. J. Neurophysiol. 66, 29–39. 25. Oppenheimer S. M. and Cechetto D. F. (1990) Cardiac chronotropic organization of the rat insular cortex. Brain Res. 533, 66–72. 26. Oppenheimer S. M., Wilson J. X., Guiraudon C. and Cechetto D. F. (1991) Insular cortex stimulation produces lethal cardiac arrhythmias: a mechanism of sudden death? Brain Res. 550, 115–121. 27. Robinson C. J. and Burton H. (1980) Organization of somatosensory receptive fields in cortical areas 7b, retroinsula, postauditory and granular insula of M. fascicularis. J. comp. Neurol. 192, 69–92. 28. Robinson C. J. and Burton H. (1980) Somatic submodality distribution within the second somatosensory (SII), 7b, retroinsular, postauditory, and granular insular cortical areas of M. fascicularis. J. comp. Neurol. 192, 93–108. 29. Ruggiero D. A., Mraovitch S., Granata A. R., Anwar M. and Reis D. J. (1987) A role of insular cortex in cardiovascular function. J. comp. Neurol. 257, 189–207. 30. Schneider R. J., Friedman D. P. and Mishkin M. (1993) A modality-specific somatosensory area within the insula of the rhesus monkey. Brain Res. 621, 116–120. 31. Szabo J. and Cowan W. M. (1984) A stereotaxic atlas of the brain of the cynomolgus monkey (Macaca fascicularis). J. comp. Neurol. 222, 265–300. 32. Taguchi H., Masuda T. and Yokota T. (1987) Cardiac sympathetic afferent input onto neurons in nucleus ventralis posterolateralis in cat thalamus. Brain Res. 436, 240–252.
360
Z.-H. Zhang et al.
33. Tang J., Burkey A. R. and Jasmin L. (1997) Comparative anatomy of the rostral agranular insular cortex of the rat with forebrain structures implicated in central pain. Soc. Neurosci. Abstr. 23, 440. 34. Verberne A. J. M. and Owens N. C. (1997) Cortical modulation of the cardiovascular system. Prog. Neurobiol. 54, 149–168. 35. Wall P. D. and Davis G. D. (1951) Three cerebral cortical systems affecting autonomic function. J. Neurophysiol. 14, 507–517. 36. Willis W. D., Jr (1985) The pain system. In Pain and Headache (ed. Gildenberg P. L.), pp. 1–346. Karger, Basel. 37. Yasui Y., Breder C. D., Saper C. and Cechetto D. F. (1991) Autonomic responses and efferent pathways from the insular cortex in the rat. J. comp. Neurol. 303, 355–374. 38. Zhang Z. H., Xie Y. F. and Qiao J. T. (1991) Effects of stimulation of superior colliculus on nociceptive unit discharges in parafascicular neurons and nocifensive reflex of hind limb in rat. Brain Res. 542, 248–253. 39. Zhang Z. H. and Oppenheimer S. M. (1997) Characterization, distribution and lateralization of baroreceptor-related neurons in the rat insular cortex. Brain Res. 760, 243–250. 40. Zhang Z. H. and Oppenheimer S. M. (1997) Reciprocal connections between insular cortex and thalamic baroreceptive and nociceptive convergent neurons in the rat. Soc. Neurosci. Abstr. 23, 423. 41. Zhang Z. H., Dougherty P. M. and Oppenheimer S. M. (1998) Characterization of baroreceptor-related neurons in the monkey insular cortex. Brain Res. 796, 303–306. (Accepted 9 June 1999)
Pergamon PII: S0306-4522(99)00339-5
Copyright q 1999 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306-4522/99 $20.00+0.00
MONKEY INSULAR CORTEX NEURONS RESPOND TO BARORECEPTIVE AND SOMATOSENSORY CONVERGENT INPUTS Z.-H. ZHANG,* P. M. DOUGHERTY† and S. M. OPPENHEIMER*‡ *Laboratory of Neurocardiology, Cerebrovascular Division, Department of Neurology and †Departments of Neurosurgery and Neuroscience, The Johns Hopkins University School of Medicine, Baltimore, MD 21287-7585, U.S.A.
Abstract—To investigate possible convergence of autonomic and somatosensory input in the insula of the non-human primate, extracellular single-unit recordings were obtained from 81 neurons (43 insular and 38 in surrounding cortex) during application of cutaneous nociceptive stimuli (pinch) and baroreceptor challenge in six anesthetized monkeys (Macaca fascicularis). All cells were also tested with light touch (brush) stimulation. Twenty-six units responded to blood pressure changes; 20 (80%) were identified within the insula (P , 0.001). The majority of these insular units (16/20) also responded to nociceptive pinch (convergent units). More units responsive to changes in blood pressure (unimodal and convergent) were found in the right (18/29, 62%) than in the left insular cortex (2/14, 14%)(P 0.004). Twenty-nine insular neurons responded to nociceptive stimuli; 16 of these were convergent units and 13 showed unimodal responses to somatosensory stimuli alone. These cells had wide bilateral receptive fields including face, hand, foot and tail. Ten insular neurons were unresponsive to both sets of stimuli (non-responsive cells); significantly more of these cells (28/38) were identified in extrainsular locations (P , 0.01). We suggest that the primate insular cortex may be involved in the integration of cardiovascular function with somatosensory (principally nociceptive) input. This view supports the emerging role of the insular cortex as an important forebrain site of viscerosomatosensory regulation with clinical implications for cardiovascular regulation under conditions of stress and arousal. q 1999 IBRO. Published by Elsevier Science Ltd. Key words: extracellular recordings, insular cortex, pain, cardiovascular regulation.
the Study of Pain and the NIH guide for the care and use of laboratory animals. Six adult male monkeys (Macaca fascicularis) weighing 4.5–7.5 kg were tranquilized with ketamine (10 mg/kg, i.m.) during transportation to the laboratory and catheterization. The right radial artery was cannulated with PE-90 tubing filled with heparinized saline (10 units/ml) for blood pressure recording. The right cephalic vein was cannulated with an i.v. catheter for infusion of anesthetics and drugs. Surgical anesthesia was induced by an intravenous dose of pentobarbital (25 mg/kg) and later maintained by intravenous infusion of pentobarbital (5 mg/kg/h). Once anesthesia was deep, the monkeys were paralysed with pancuronium bromide (0.1 mg/kg) and artificially ventilated. Throughout the surgical procedures the level of anesthesia was periodically tested and adjusted as necessary by examining pupillary size, and continuously monitored by recording expired CO2, blood pressure and heart rate. End-tidal CO2 was kept between 3.5% and 4.5% by adjustment of respiration rate and depth, and core temperature was maintained at 37–388C by a servo-controlled heating blanket. The animals were placed in a stereotaxic frame and a partial craniectomy was made on both sides for exposure of the insular cortex.
In many species, the insular cortex is an important cortical site of cardiovascular representation. 2,10,25,29,34,41 In the rat, both electrical and chemical stimulation of the insula caudal to the crossing of the anterior commissure produces cardiac chronotropic responses and blood pressure changes. 25,37 Prolonged insular stimulation in rats generates lethal cardiac arrhythmias and structural changes in cardiac muscle. 26 Previous electrical stimulation studies in the monkey have suggested similar insular involvement in cardiovascular control but lacked the precision of contemporary methods. 20,35 Anatomical evidence in primates has implicated the insula in the processing of somatosensory innocuous and nociceptive information. 17 In addition, functional imaging studies in humans using positron emission tomography have demonstrated insular activation with somatosensory (especially thermal and vibrotactile) stimuli. 8,13,15 The goal of this study was to investigate whether primate baroreceptor responsive insular neurons also respond to somatic stimuli, as has recently been shown in the rat insular cortex 19 and thalamus. 40 Such convergence at cortical levels could further support the emerging role of the insula as a site of integration of pain and autonomic function.
Recording procedures Extracellular recordings were made from spontaneously firing neurons in the insular cortex and surrounding regions with carbon filament-glass microelectrodes (impedance 1–2 MV at 1 kHz). Coordinates for the regions of study (AP 11–20 mm; ML 15–18 mm; DEPTH 10–20 mm from cortical surface) were based on the stereotaxic atlas of Szabo and Cowan 31 adjusted for body weight. Singleunit activity, electrocardiogram and arterial pressure were amplified and displayed on an oscilloscope (Hitachi VC 6155, Japan) and fed to a computer running a neurophysiological data acquisition and analysis package (Datawave Systems, Discovery package, Version 4.0, Thornton, CO, U.S.A.). Spike size and configuration were continuously monitored to confirm that activity of the same cell was recorded throughout the experiment. Single-unit activity was discriminated from background and stored on computer for later off-line analysis.
EXPERIMENTAL PROCEDURES
Surgical procedures All procedures were approved by the Institutional Animal Care and Use Committee of the Johns Hopkins University and are consistent with the guidelines of the International Association for ‡To whom correspondence should be addressed. Abbreviations: PB, pinch and brush; PE, phenylephrine hydrochloride; PET, positron emission tomography; PR, pinch responsive; SE, sympathoexcitatory; SI, sympathoinhibitory; SNP, sodium nitroprusside; VPL, ventroposterior lateral nucleus.
Baroreceptor challenge and cutaneous stimuli Following unit isolation and stabilization, 3 min of spontaneous 351
352
Z.-H. Zhang et al.
the recording sites were identified with reference to a measured scale taking into account the magnification and the distance between the recording site and the bottom of the microelectrode track. Correction for brain shrinkage was made to ensure precise histology by measuring the depth of the track in the pre- and postfixed states. Tissue distortion interfering with track reconstruction was rarely encountered. Significant brain edema was not generally seen. In all cases (except one) the entire track could be seen in at least one section, making reconstruction feasible (Fig. 1). However, in one monkey, tissue was distorted owing to brain edema. Identification of the bottom of the track was not impaired and correction for the distortion was possible by careful reconstruction of the track in sequential sections. Data analysis The stored digital records of unit activity before and after baroreceptor challenge and somatosensory stimulation were retrieved and analysed off-line. Baroreceptive responses were analysed by comparison of discharge rates for 90–120 s after drug injection compared with a similar period before injection. Mechanoreceptor responses were analysed by comparison of discharge rates in each separate 10 s segment of stimulus application to a 10 s segment of pre-stimulus activity. All values were expressed as the mean ^ S.E.M. Changes in firing rates for individual cells . ^ 20% were considered significant. 1,38,39,41 Cells with such changes were then categorized into groups according to their responses to PE, SNP, pinch and brush. Changes in their mean firing rate for the periods before and during challenge with baroreceptor and somatosensory stimuli were analysed statistically using Student’s t-test. The distributions of responsive neurons were assessed for intrainsular or extrainsular localization and lateralization with Fisher’s exact test. P , 0.05 indicated statistical significance. Fig. 1. Photomicrograph of a Thionine-stained section showing the track (indicated by the arrow) through the dysgranular responsive laminae of the insular cortex. Scale bar 5 mm.
baseline activity was recorded. Arterial baroreceptors were then challenged by bolus intravenous administration of the pressor agent phenylephrine hydrochloride (PE, 10 mg/kg, Sigma Chemical Co., St Louis, U.S.A.) and the depressor drug sodium nitroprusside (SNP, 5 mg/kg, Sigma Chemical Co., St Louis, U.S.A.) in a volume of 0.2 ml 0.9% saline, followed by 0.2 ml 0.9% saline to flush the dead space of the tubing (dead space volume 0.1 ml). These doses induced ^20–40 mmHg change in arterial pressure. Injection of 0.2 ml 0.9% saline was used as a control procedure. These injections were given in random sequence. After characterization of unit responses to baroreceptor activation and re-establishment of baseline firing conditions, a further 2 min of spontaneous baseline firing activity was collected. The responses of cells to a series of mechanical stimuli applied to the selected skin sites were then determined. Cutaneous mechanoreceptor stimuli included application of a camel hair brush and arterial clip to seven areas of the body (cheeks, hands, feet and tail). Clip pressure was standardized using a spring mechanism. When applied to the human skin, the clip elicited pain in a reproducible fashion. Each stimulus set began with recording of background activity followed by the brush (back and forth brushing) applied to site 1 (randomized) for 10 s. Following a 10–20 s interval, or until baseline firing rate was re-established following removal of the stimulus, the brush was then applied to site 2. This continued until the brush had been applied to all sites. The same sequence was then repeated for the pinch stimulus. Care was taken to ensure that the brush and pinch stimuli were delivered in a stereotyped manner on each occasion. Histology At the conclusion of each experiment, the last recording site was marked by the insertion of a steel electrode for placement of an electrolytic lesion (100 mA d.c. anodal current, 10 s duration). Other cell sites were identified with respect to this reference point. The animals were killed by an overdose of pentobarbital, and the brain was removed and postfixed in 10% formalin for at least 10 days. Frontal sections of 50 mm thickness were cut with a cryostat microtome. Sections were mounted on gelatin-coated glass slides, dried overnight, dehydrated in a series of graded alcohols and then stained with 0.125% Thionin solution. The slide was projected using a Zeiss camera lucida, and
RESULTS
Neuronal responses to baroreceptor and mechanoreceptor stimuli were investigated in 81 units, 43 within the insula and 38 in adjacent cortex. As the investigation was primarily designed to investigate the presence of convergent cardiovascular-nociceptive cells within the insula, no attempt was made to investigate subcortical areas including the claustrum and basal ganglia. The insular cortex was defined as that part of the cerebrum that overlies the claustrum. In primates this region lies buried beneath the frontoparietal and temporal opercula. Three cytoarchitectural divisions or fields of the insula occur in the monkey. 2,21 These include a rostroventral agranular field, a transitional dysgranular field and, posteriorly, a granular field. Cells were categorized in accordance with their firing responses to baroreceptor activation by PE and SNP-induced blood pressure changes 3,39,41 and by responses to somatic stimuli. 1,18,36 Convergent units responded to both blood pressure changes and somatic stimulation. Neuronal response patterns to baroreceptor activation A total of 26 units (20 in the insula and six in the adjacent cortex) responded to baroreceptor challenge with PE and SNP. According to their response to PE and SNP, two types of unit were categorized as reported previously in the rat 3,9,39 and monkey. 41 Sympathoexcitatory (SE) units were identified by a significantly decreased firing rate following PE-induced increase in blood pressure. The firing rate of these units increased with SNP-related reduction in blood pressure (Fig. 2A, B). Sympathoinhibitory (SI) units showed an increased firing rate with PE injection and decreased firing following SNP administration (Fig. 3A, B). The classification of cells as either SE or SI units is a convention 3,39,41 and does not necessarily mean that the physiological effect of these cells is either to increase or decrease cardiovascular
Fig. 2. Firing patterns of a representative SE nociceptive convergent unit in the monkey insula. Firing rate decreases with PE (10 mg/kg) induced blood pressure rises (A) and increases with SNP (5 mg/kg) induced falls in blood pressure (B). Panel (C) shows absence of response to the application of brush stimulation at seven sites. Application of pinch to these seven sites increases discharge rate (D). In each panel: Top trace shows blood pressure (BP); middle trace shows the footpad marker of seven brush (C) and pinch (D) sites; bottom trace shows the frequency histogram of neural firing (binwidth 1 s). The right column shows the superimposed traces of neural spikes showing consistency of amplitude and shape throughout the recording. Asterisks show significant increases in firing rate during pinch stimuli; the corresponding responsive sites are marked by black dots on the monkey diagram (D). The arrows indicate drug administration (A and B).
354
Z.-H. Zhang et al.
Fig. 3. A representative response of a SI nociceptive convergent unit in monkey insular cortex showing firing rate increases after PE (10 mg/kg) injection (A) and decreases following SNP (5 mg/kg) administration (B). No significant response to brush correlating with stimulus application (C). Significant increases in firing rate to pinch in all seven sites (D). Display and abbreviations are the same as for Fig. 1.
Convergent baroreceptor and somatic neurons in monkey insular cortex
sympathetic tone. Non-responsive units showed no response to either baroreceptor or somatic activation (Fig. 4). Equal volume injections of saline had no effect on discharge rates in these same cells. The majority of insular cortex responsive cells showed SE responses to blood pressure changes (14/20, 70%) (Table 1). SI responses (6/20, 30%) were relatively infrequently encountered in the insula (Table 1). The mean firing rates of SE and SI units are shown in Fig. 5. For SE units the mean onset latency was 7.9 ^ 2.6 s (range 0–20 s) for PE injections and 10.4 ^ 3.2 s (range 0–22 s) for SNP injections. The mean duration of the SE response to PE injections was 102 ^ 14 s (range 20–180 s), and for SNP the mean duration of the response was 94 ^ 13 s (range 20– 160 s). For SI units, the mean onset latency for PE injections was 8.6 ^ 2.4 s (range 0–15 s) and 11.2 ^ 2.7 s (range 0–18s) for SNP injections. The response duration to PE was 94 ^ 16 s (range 20–150 s) and 97 ^ 16 s (range 30–170 s) for SNP. Neuronal response patterns to somatic stimuli Thirty-seven units (29 in the insula and eight outside the insula) showed a significant change in firing rate with mechanical stimulation of the skin. Responses showed clear onset and offset following the 10 s application of somatic stimuli. The mean firing frequency before and during somatic stimulation for each group is shown in Fig. 5. Most cells (n 31) showed increased discharges to nociceptive stimulation (on-cell) while the remainder (n 6) showed decreased firing (off-cell). Thirty-one showed responses to pinch only [26 on-cells, five off-cells; pinchresponsive (PR) units]; six units responded to both pinch and brush (PB units); no units were observed to respond to brush alone (Fig. 5, Table 2). Nociceptive units (PR and PB cells) had relatively wide receptive fields: 19/37 (51%) responded to stimulation applied to five or more sites throughout the upper and lower body; 6/37 (16%) cells responded to contralateral somatic sites. Five cells responded to bilateral “upper” body stimuli and three cells to bilateral “lower” body stimuli. Four neurons responded only to ipsilateral stimuli applied to more than one site. Pinch induced a very slight decrease in blood pressure (,10 mmHg) but a marked increase in firing rate (Figs 2, 3). These blood pressure responses were highly variable in their occurrence (Figs 2, 3). The effect of SNP injection was to induce a much greater fall in blood pressure in SE cells than that which occurred with pinch, but this was accompanied by a much smaller increase in firing rate. For SI cells, where pinch induced a small change in blood pressure, the effect on neural firing was opposite to that induced by SNP (Fig. 3). Categorization of convergent neuronal responses Twenty units responded to both baroreceptor challenge and somatic mechanical stimuli (convergent units). Sixteen of 29 insular cells were convergent units; four extrainsular cells showed such convergent responses. Most of the insular convergent units (11/16, 69%) showed sympathoexcitatory responses to baroreceptor activation (Table 2, Fig. 6). Five SI convergent units were found within the insular cortex (Table 2, Fig. 6). Fifteen of 20 convergent cells responded
355
only to pinch stimuli (PR cells); 5/20 convergent units responded to both pinch and brush stimulation (PB cells); no convergent units showed responses to brush alone (Table 2, Fig. 5). Thirty-eight non-responsive neurons responded to neither baroreceptor challenge nor somatic stimuli (Table 2, Figs 5, 6). Localization of responsive and non-responsive units The majority of insular neurons (33/43, 77%) responded to baroreceptor and/or somatic activation. Fewer such responses were seen in the adjacent cortex (10/38, 26%) (Fig. 6, Table 1). More units responsive to changes in blood pressure (unimodal and convergent) were found in the right (18/29, 62%) than in the left insular cortex (2/14, 14%)(P 0.004, Fig. 6, Table 1). There was no clear localization of responsive cell types in the rostrocaudal insular axis. A significantly greater number of cells unresponsive to either blood pressure changes or cutaneous mechanoreceptor stimuli was found in extrainsular (28/38, 74%) compared with intrainsular locations (10/43, 23%) (P , 0.01, Table 1, Fig. 6). Responsive insular neurons were chiefly identified within the midinsular dysgranular region in layers II, III and V/VI (Figs 1, 6). Few responsive units were identified in the agranular insular cortex. This may reflect the limited sampling of the agranular region. More anterior and posterior insular regions were not systematically sampled. Responsive neurons outside the insula were encountered in the left mid-frontoparietal operculum adjacent to the insula (two convergent, two SI and one PR unit) and left temporal operculum (one PB unit), and right superior/mid-temporal cortex (two convergent and two PR units) (Fig. 6, Table 1). DISCUSSION
Response patterns of insular convergent neurons This study shows that cells convergent for baroreceptive and somatosensory input can be identified within the insular cortex of the anesthetized monkey. As previously reported in the rat and monkey, the majority of insular neurons responsive to blood pressure changes were SE units. 39,41 Convergent units generally responded to nociceptive stimuli. The data showed the possibility of lateralization of units responsive to blood pressure changes within the primate insula. However, this may reflect sampling bias as fewer left anterior insular regions were sampled in this study. The level at which convergence of baroreceptive and nociceptive input occurs initially within the CNS cannot be determined by this study. However, we have recently demonstrated such convergence within the posterior thalamus of the rat in units which project monosynaptically to the insula, 40 suggesting that convergence is likely to occur at subcortical sites in the monkey. Visceral (non-cardiovascular) and somatic convergent neurons have been demonstrated within the ventroposterior lateral nucleus (VPL) of the monkey 4 and cat thalamus. 32 A recent study suggested that an inhibitory interaction occurs between visceral and somatic inputs in the thalamic VPL of the squirrel monkey. 5 This implies that either somatic conditioning stimuli inhibit visceral responses, or vice versa. Similar interactions have been identified at different levels of the CNS in other species; 11,12,23,24 however, this has not yet been demonstrated in the primate insular cortex. We believe for several reasons that a distinction can be
356
Z.-H. Zhang et al.
Fig. 4. Representation of a non-responsive unit in the extrainsular frontoparietal operculum showing no significant change in firing rate following blood pressure alteration by PE and SNP (A and B). No response to brush and pinch stimuli (C, D). Display and abbreviations are the same as for Fig. 1.
Convergent baroreceptor and somatic neurons in monkey insular cortex Table 1. Distribution of baroreceptive, somatosensory, convergent and nonresponsive cells within and outside the insula
SE cell SI cell PR and PB cell Convergent cell Non-responsive cell Total
Right in IC
Left in IC
Right out IC
Left out IC
Total
3 1 7 14 4 29
0 0 6 2 6 14
0 0 2 2 13 17
0 2 2 2 15 21
3 3 17 20 38 81
SE and SI cells respond only to baroreceptor stimulation with PE and SNP. PR cells respond to pinch only and PB cells respond to pinch and brush. Non-responsive cells respond to neither baroreceptor nor pinch and brush stimulation. IC, insular cortex.
made in the firing of convergent insular neurons to nociceptive stimuli. The responses are not simply due to the slight changes in blood pressure induced by pinch. Firstly, the nociceptive responses have a clear onset and offset with application of pinch stimuli and do not correspond to the somewhat delayed blood pressure responses to pinch (Figs 2, 3). Secondly, the majority of cells showed an increase in firing rate with SNP-induced blood pressure decrease (SE cells). Although pinch also induced a decrease in blood pressure, this was minimal compared with the SNP-induced blood pressure changes, but was associated with a much greater increase in firing than achieved by the larger SNP-induced decrease in blood pressure (Fig. 2). In the case of SI cells, pinch induced a decrease in blood pressure but an increase in firing, opposite to the effects of SNP on the cell (Fig. 3). Finally, blood pressure responses to pinch were inconsistent; responses to pinch were often obtained in the absence of any blood pressure effect (Figs 2, 3). Whether convergent and non-responsive cells in the
357
Table 2. Classification of baroreceptive and somatosensory units Somatosensory responses
PR PB NR Total
Baroreceptive responses SE SI NR 11 3 3 17
4 2 3 9
16 1 38 55
Total 31 6 44 81
NR, non-responsive units.
primate insula respond to other sensory modalities such as taste or gastric mechanical input, as has been shown for rat anterior insular neurons by Cechetto and Saper, 9 was not the objective of this investigation, which focussed on the demonstration of nociceptive–baroreceptive convergence. In theory, specificity of neuronal response to baroreceptor input may be confirmed by baroreceptor denervation. This would counter arguments that responses to blood pressure manipulations were produced by stimulation of mechanoreceptors in the arterial wall or adjacent structures rather than by stimulation of baroreceptors in the aortic arch, heart and carotid sinus. In practice, however, baroreceptor denervation in the monkey also produces an unstable preparation. The cardiovascular status of the animals in these two widely differing situations (pre- and post-denervation) would no longer be comparable, confounding comparison of cellular responses between these states. We argue that there is specificity of baroreceptor response on the basis of the reproducible manner in which these cells responded to standard techniques of baroreceptor activation with PE and SNP, as we have already shown for insular baroreceptive neurons in the rat 39 and monkey. 41 Other studies have indicated that similar stimuli induce their effect on units in the brain by activation of peripheral visceral receptors. 9
Fig. 5. Statistical analysis of firing rates of cells divided into groups according to their responses to baroreceptor and mechanical stimulation. Data are shown for all 81 cells in the study. SE convergent cells are characterized as in Fig. 1; SI convergent cells are defined as in Fig. 2. PR and PB cells respond to pinch (PR cells) or pinch and brush stimuli (PB cells) but not to baroreceptor challenge with PE and SNP; SE and SI cells respond only to baroreceptor challenge with PE and SNP (as defined in Figs 1 and 2) but not to somatosensory stimuli. Statistical analysis was conducted for changes in unit firing rates to somatosensory stimuli application (brush and pinch) at all responsive sites. On or off cell responses to pinch within each group are shown separately (Pinch-on; pinch-off). No off responses to brush stimulation were encountered. No statistical testing was conducted on unimodal SE (n 3) and SI cells (n 3) in response to SNP and PE challenge, because of the small number of cells with these characteristics. In each group, the hatched bars represent baseline firing rates before drug injection or mechanoreceptor stimuli (control); the solid bars reflect the changes in firing rate after drug infusion or during mechanical stimulation. *P , 0.05; **P , 0.01; ***P , 0.001 (Student’s t-test for paired data).
358
Z.-H. Zhang et al.
Functional organization of the monkey insular cortex Cardiovascular responses have previously been shown in earlier stimulation studies of the monkey insula. 20,35 However,
care must be taken in the interpretation of these data because responses were elicited with large constant voltage stimuli of long duration (2–10 ms) using large surface stimulating electrodes. Consequently, the specificity of the responses and their localization to the stimulation site may be questionable. Previous studies indicated that approximately 70% of neurons in the granular posterior insular cortex respond to non-noxious somatic inputs in anesthetized cynomolgus monkeys. 27,28 Only 24 neurons were investigated in this region, of which 25% also responded to nociceptive stimuli. We did not identify any cells that responded solely to nonnociceptive stimuli. However, we explored the dysgranular mid-insula and did not sample the very posterior regions of the insular cortex studied by Robinson and Burton. 27,28 This methodological difference may explain this discrepancy. Sixty per cent of insular neurons in the Robinson and Burton study responded to bilateral somatic stimulation, consistent with our findings for PR and PB cells. Similar results have been found in unanesthetized macaques. 30 The monkey insula receives projections from the posterior part of the thalamic ventral medial nucleus, 14 which in turn receives input from spinal dorsal horn lamina I neurons, a site of termination of nociceptive primary afferents. Burkey et al. 6 recently found that microinjection of morphine into an anatomically defined opioid-responsive site in the rostral agranular insular cortex of the rat significantly reduced the behavioral response to noxious stimuli. Their more recent data 7 indicated that a descending tonic inhibitory dopaminergic pathway from the insula can modulate nociception. Anatomical studies suggest from a comparative consideration of connectivity that the rostral agranular insular cortex in the rat is homologous to more posterior areas in the primate insular cortex which have been implicated in the development of post-stroke pain syndromes. 33 It is worthy of note that we found that some cells in the area adjoining the insular cortex also receive baroreceptor and/or nociceptive inputs. No overall conclusion can be reached, owing to limited sampling of these areas. Recent functional imaging studies show involvement of the frontoparietal operculum in nociceptive processing in the rat. 22 Our studies in the rat and monkey also showed that a small percentage of neurons adjacent to the insular cortex (most in the parietal operculum) responds to blood pressure change. 39,41 However, further studies are needed to clarify this issue. Nociceptive input into the insular cortex has also been shown using positron emission tomography (PET). Increased insular cerebral blood flow occurred in patients complaining of atypical facial pain when painful but not when non-painful thermal stimulation was applied to the back of the contralateral hand. 16 Interestingly, normal controls did not show this increased insular activation, suggesting the involvement of the insula in the affective component of pain. Likewise, functional imaging studies in humans show strong activation of the mid/anterior insula by noxious and innocuous thermal Fig. 6. Distribution of convergent (B), SE (X), SI (K), nociceptive (V), and non-responsive (W) units within the monkey insula and surrounding regions. Numbers indicate the distance in mm anterior to the interaural zero point. Amyg, amygdala; CA, anterior commissure; Cd, caudate nucleus; Cho, optic chiasm; Cl, claustrum; Gp, globus pallidus; Hyp, hypothalamus; Put, putamen; TO, optic tract; VA, ventral anterior nucleus of the thalamus; VLm, ventral lateral nucleus of the thalamus, medial part; VLo, ventral lateral nucleus of the thalamus, oral part. Figure templates are modified from the atlas of Szabo and Cowan. 31
Convergent baroreceptor and somatic neurons in monkey insular cortex
stimuli. 8,15 It has been hypothesized by these investigators that the insula is concerned with the conscious perception and discrimination of the affective qualities of both thermal and noxious stimuli and may be involved in the generation of illusory centrally mediated sensations such as phantom limb pain. Coghill and colleagues 13 used PET to localize differential insular activation with vibrotactile stimulation compared with noxious thermal stimuli applied to the left forearm. Nociceptive activation of the insula was noted in the principally contralateral anterior regions, although some ipsilateral responses were apparent. Vibrotactile responses were principally confined to the posterior insula. As in our study, they concluded that there was significant input of nociceptive information into the insula and suggested possible involvement of the dysgranular insula consistent with our findings. However, we did not find cells solely responsive to non-nociceptive inputs. This discrepancy may relate to the use of
359
anesthesia in this study, species differences and the different stimuli used to elicit tactile responses. CONCLUSIONS
Our data indicate significant convergence of somatosensory (principally nociceptive) and baroreceptive input on the same neurons in the primate insula. This region has recently been shown to be involved with the affective and perceptive elements of pain processing in human functional imaging studies. There is much recent evidence indicating that the insula also has an important role in the efferent control of cardiovascular function. Taken together, these data indicate that the insular cortex may co-ordinate the autonomic aspects of central pain processing preparing the animal for appropriate enhancement or suppression of cardiovascular function as the environmental circumstances dictate.
REFERENCES
1. Andersen E. and Dafny N. (1983) Dorsal raphe stimulation reduces responses of parafascicular neurons to noxious stimulation. Pain 15, 323–331. 2. Augustine J. R. (1996) Circuitry and functional aspects of the insular lobe in primates including humans. Brain Res. Rev. 22, 229–244. 3. Barman S. M. (1990) Descending projections of hypothalamic neurons with sympathetic nerve-related activity. J. Neurophysiol. 64, 1019–1032. 4. Bruggemann J., Shi T. and Apkarian A. V. (1994) Squirrel monkey lateral thalamus. II. Viscerosomatic convergent representation of urinary bladder, colon, and esophagus. J. Neurosci. 14, 6796–6814. 5. Bruggemann J., Shi T. and Apkarian A. V. (1998) Viscerosomatic interactions in the thalamic ventral posterolateral nucleus (VPL) of the squirrel monkey. Brain Res. 787, 269–276. 6. Burkey A. R., Carstens E., Wenniger J. J., Tang J. W. and Jasmin L. (1996) An opioidergic cortical antinociception triggering site in the agranular insular cortex of the rat that contributes to morphine antinociception. J. Neurosci. 16, 6612–6623. 7. Burkey A. R., Carstens E., Wenniger J. J. and Jasmin L. (1997) Antinociception and descending inhibition from dopamine in the rostral agranular insular cortex of the rat. Soc. Neurosci. Abstr. 23, 160. 8. Casey K. L., Minoshima S., Morrow T. J. and Koeppe R. A. (1996) Comparison of human cerebral activation patterns during cutaneous warmth, heat pain, and deep cold pain. J. Neurophysiol. 76, 571–581. 9. Cechetto D. F. and Saper C. B. (1987) Evidence for a viscerotopic sensory representation in the cortex and thalamus in the rat. J. comp. Neurol. 262, 27–45. 10. Cechetto D. F. (1994) Identification of a cortical site for stress-induced cardiovascular dysfunction. Integr. physiol. behav. Sci. 29, 362–373. 11. Chernigovsky V. N., Musyashchikova S. S. and Sinyaya M. S. (1978) Interaction of visceral and somatic afferent systems in the cerebral cortex. Neurosci. behav. Physiol. 9, 273–282. 12. Coffin B., Azpiroz F. and Malagelada J. R. (1994) Somatic stimulation reduces perception of gut distention in humans. Gastroenterology 107, 1636–1642. 13. Coghill R. C., Talbot J. D., Evans A. C., Meyer E., Gjedde A., Bushell M. C. and Duncan G. H. (1994) Distributed processing of pain and vibration by the human brain. J. Neurosci. 14, 4095–4108. 14. Craig A. D. (1996) Supraspinal projections of lamina I neurons. In Forebrain Processing of Pain (ed. Besson J.-M.), pp. 13–25. John Libbey, Paris. 15. Craig A. D., Reiman E. M., Evans A. and Bushnell M. C. (1996) Functional imaging of an illusion of pain. Nature 384, 258–260. 16. Derbyshire S. W. G., Jones A. K. P., Devani P., Friston K. J., Feinmann C., Harris M., Pearce S., Watson J. D. G. and Frackowiak R. S. J. (1994) Cerebral responses to pain in patients with atypical facial pain measured by positron emission tomography. J. Neurol. Neurosurg. Psychiat. 57, 1166–1172. 17. Dostrovsky J. O. and Craig A. D. (1996) Nociceptive neurons in primate insular cortex. Soc. Neurosci. Abstr. 22, 111. 18. Dougherty P. M., Palecek J., Paleckova V., Sorkin L. S. and Willis W. D. (1992) The role of NMDA and non-NMDA excitatory amino acid receptors in the excitation of primate spinothalamic tract neurons by mechanical, chemical, thermal, and electrical stimuli. J. Neurosci. 12, 3025–3041. 19. Hanamori T., Kunitake T. and Kannan H. (1998) Responses of neurons in the insular cortex to gustatory, visceral, and nociceptive stimuli in rats. J. Neurophysiol. 79, 2535–2545. 20. Hoffman B. L. and Rasmussen T. (1953) Stimulation studies of insular cortex of Macaca mulatta. J. Neurophysiol. 16, 343–351. 21. Mesulam M.-M. and Mufson E. J. (1982) Insula of the old world monkey. I: Architectonics in the insulo-orbito-temporal component of the paralimbic brain. J. comp. Neurol. 212, 1–22. 22. Morrow T. J., Paulson P. E., Danneman P. J. and Casey K. L. (1998) Regional changes in forebrain activation during the early and late phase of formalin nociception: analysis using cerebral blood flow in the rat. Pain 75, 355–365. 23. Ness T. J. and Gebhart G. F. (1991) Interactions between visceral and cutaneous nociception in the rat. I. Noxious cutaneous stimuli inhibit visceral nociceptive neurons and reflexes. J. Neurophysiol. 66, 20–28. 24. Ness T. J. and Gebhart G. F. (1991) Interactions between visceral and cutaneous nociception in the rat. II. Noxious visceral stimuli inhibit cutaneous nociceptive neurons and reflexes. J. Neurophysiol. 66, 29–39. 25. Oppenheimer S. M. and Cechetto D. F. (1990) Cardiac chronotropic organization of the rat insular cortex. Brain Res. 533, 66–72. 26. Oppenheimer S. M., Wilson J. X., Guiraudon C. and Cechetto D. F. (1991) Insular cortex stimulation produces lethal cardiac arrhythmias: a mechanism of sudden death? Brain Res. 550, 115–121. 27. Robinson C. J. and Burton H. (1980) Organization of somatosensory receptive fields in cortical areas 7b, retroinsula, postauditory and granular insula of M. fascicularis. J. comp. Neurol. 192, 69–92. 28. Robinson C. J. and Burton H. (1980) Somatic submodality distribution within the second somatosensory (SII), 7b, retroinsular, postauditory, and granular insular cortical areas of M. fascicularis. J. comp. Neurol. 192, 93–108. 29. Ruggiero D. A., Mraovitch S., Granata A. R., Anwar M. and Reis D. J. (1987) A role of insular cortex in cardiovascular function. J. comp. Neurol. 257, 189–207. 30. Schneider R. J., Friedman D. P. and Mishkin M. (1993) A modality-specific somatosensory area within the insula of the rhesus monkey. Brain Res. 621, 116–120. 31. Szabo J. and Cowan W. M. (1984) A stereotaxic atlas of the brain of the cynomolgus monkey (Macaca fascicularis). J. comp. Neurol. 222, 265–300. 32. Taguchi H., Masuda T. and Yokota T. (1987) Cardiac sympathetic afferent input onto neurons in nucleus ventralis posterolateralis in cat thalamus. Brain Res. 436, 240–252.
360
Z.-H. Zhang et al.
33. Tang J., Burkey A. R. and Jasmin L. (1997) Comparative anatomy of the rostral agranular insular cortex of the rat with forebrain structures implicated in central pain. Soc. Neurosci. Abstr. 23, 440. 34. Verberne A. J. M. and Owens N. C. (1997) Cortical modulation of the cardiovascular system. Prog. Neurobiol. 54, 149–168. 35. Wall P. D. and Davis G. D. (1951) Three cerebral cortical systems affecting autonomic function. J. Neurophysiol. 14, 507–517. 36. Willis W. D., Jr (1985) The pain system. In Pain and Headache (ed. Gildenberg P. L.), pp. 1–346. Karger, Basel. 37. Yasui Y., Breder C. D., Saper C. and Cechetto D. F. (1991) Autonomic responses and efferent pathways from the insular cortex in the rat. J. comp. Neurol. 303, 355–374. 38. Zhang Z. H., Xie Y. F. and Qiao J. T. (1991) Effects of stimulation of superior colliculus on nociceptive unit discharges in parafascicular neurons and nocifensive reflex of hind limb in rat. Brain Res. 542, 248–253. 39. Zhang Z. H. and Oppenheimer S. M. (1997) Characterization, distribution and lateralization of baroreceptor-related neurons in the rat insular cortex. Brain Res. 760, 243–250. 40. Zhang Z. H. and Oppenheimer S. M. (1997) Reciprocal connections between insular cortex and thalamic baroreceptive and nociceptive convergent neurons in the rat. Soc. Neurosci. Abstr. 23, 423. 41. Zhang Z. H., Dougherty P. M. and Oppenheimer S. M. (1998) Characterization of baroreceptor-related neurons in the monkey insular cortex. Brain Res. 796, 303–306. (Accepted 9 June 1999)